IRC logs for #baserock for Monday, 2017-07-10

*** gtristan has quit IRC04:42
*** gtristan has joined #baserock05:15
*** paulwaters_ has joined #baserock07:38
*** toscalix has joined #baserock08:02
*** jonathanmaw has joined #baserock08:03
*** rdale has joined #baserock08:25
*** bruce has joined #baserock08:35
*** bruce is now known as Guest9563608:35
*** noisecell has joined #baserock08:36
gitlab-br-botdefinitions: merge request (staging/sam/gcc7.1-bootstrap-fix->master: Fix build issues in master after GCC 7.1 upgrade) #39 changed state ("opened"): https://gitlab.com/baserock/definitions/merge_requests/3908:51
*** tiagogomes has joined #baserock08:57
*** CTtpollard has joined #baserock09:03
*** edcragg has joined #baserock09:10
gitlab-br-botybd: merge request (jonathan/rpm-export-version-release-package->master: rpm: Write version and release on a per-package basis) #347 changed state ("opened"): https://gitlab.com/baserock/ybd/merge_requests/34709:14
gitlab-br-botybd: merge request (benbrown/lockfiles->master: Fix file locking errors with higher number of instances) #371 changed state ("opened"): https://gitlab.com/baserock/ybd/merge_requests/37109:31
*** locallycompact has joined #baserock09:32
*** lc_ has joined #baserock09:33
*** locallycompact has quit IRC09:37
*** ssam2 has joined #baserock09:43
*** ChanServ sets mode: +v ssam209:43
ssam2here's a puzzle then10:06
ssam2elfutils fails to build inside YBD, but succeeds if I run the build manually in the build chroot10:06
ssam2seems that the difference PATH is the key10:06
ssam2however... wtf10:06
ssam2seems that using /usr/lib/ccache/gcc instead of /sbin/gcc causes an implicit-fallthrough compile warning to appear (which then fails the build due to -Werror)10:08
ssam2but ... *its the same source code*10:08
noisecellssam2, different binaries?10:11
ssam2yeah, /usr/lib/ccache/gcc is just a link to /usr/bin/ccache10:12
ssam2ccache --clear doesn't help10:13
ssam2our ccache is several years old, I guess it could be bug10:16
ssam2*a bug10:16
gitlab-br-botlorries: merge request (sam/ccache-moved->master: Update ccache lorry for new GitHub repo) #17 changed state ("opened"): https://gitlab.com/baserock/lorries/merge_requests/1710:20
gitlab-br-botlorries: merge request (sam/ccache-moved->master: Update ccache lorry for new GitHub repo) #17 changed state ("opened"): https://gitlab.com/baserock/lorries/merge_requests/1710:21
gitlab-br-botlorries: merge request (sam/ccache-moved->master: Update ccache lorry for new GitHub repo) #17 changed state ("merged"): https://gitlab.com/baserock/lorries/merge_requests/1710:21
*** jonathanmaw has quit IRC10:34
*** jonathanmaw has joined #baserock10:35
*** jonathanmaw has quit IRC11:05
*** jonathanmaw has joined #baserock11:05
ssam2lorry-controller seems hopelessly broken11:26
benbrown_:\11:26
ironfootssam2: erm, sorry, will fix11:26
ssam2do you know what the problem is ?11:26
ironfoot(i thought I left it in a working state)11:27
ssam2i may have broken things further already11:27
ssam2the journal for lighttpd-lorry-controller-webapp is just a series of backtraces11:27
ssam2i tried removing the webapp.db file but that didn't fix11:27
ironfootit should be sorted now11:28
ssam2http://localhost:12765/1.0/status-html still gives a 500 error11:28
ironfootthis is now a broken DB problem (it seems)11:28
ssam2ah OK11:29
ssam2what was wrong before ?11:29
ssam2i deleted the DB but same issue11:29
ssam2"OperationalError: no such table: troves"11:30
ssam2ah, looks better now11:31
ironfootThe thing is that i found a problem i need to fix. The migrations are run everytime the database is opened11:32
ironfootThe locking context manager we have is only being used when "setting" data in the DB but not when "getting" data.11:32
* benbrown_ has noticed that11:33
ironfootBut the database is being opened and migrated also in the "getting" case, triggering some "Database is locked" issues11:33
benbrown_the migration process is supposed to be idempotent11:33
ssam2aha11:33
ironfootAt least I think that's the issue. But as always the issue may be me, and my new code :)11:33
ironfootalthough if you see now the journal (for webapp) you can see the errors I'm talking about, (with l-c master, without my code changes)11:34
ironfootbenbrown_: yep, I don't know why it's failing like that, didn't have the time to look deeper11:35
ironfootAlthough I think that making the table creation/migration run only once, when starting the server, would be an excellent idea11:36
* benbrown_ nods11:36
gitlab-br-botlorries: merge request (sam/fix-ccache->master: ccache: Remove 'maint' refspec, use defaults) #18 changed state ("opened"): https://gitlab.com/baserock/lorries/merge_requests/1811:38
gitlab-br-botlorries: merge request (sam/fix-ccache->master: ccache: Remove 'maint' refspec, use defaults) #18 changed state ("opened"): https://gitlab.com/baserock/lorries/merge_requests/1811:39
gitlab-br-botlorries: merge request (sam/fix-ccache->master: ccache: Remove 'maint' refspec, use defaults) #18 changed state ("merged"): https://gitlab.com/baserock/lorries/merge_requests/1811:39
ssam2thanks11:42
ironfootnp, and apologies11:42
ssam2no problem. lorry-controller definitely needs work ...11:43
ssam2hooray, new ccache11:43
* benbrown_ has a few things hiding in branches locally11:44
benbrown_Though I don't give myself time to finish and submit them11:44
ironfootbenbrown_: oh, what amazing features were you working on!11:55
* SotK idly wonders whether the yaml lorries support ended up in git.b.o11:56
* benbrown_ doesn't think so11:56
ironfootSotK: yep11:57
paulsher1oodbenbrown_: https://gitlab.com/baserock/ybd/merge_requests/276 - should not be merged, imo11:57
benbrown_paulsher1ood: 1st and 3rd commits seem sensible enough to me11:58
benbrown_I don't really know what the 'file:///' magic is for, tbh11:58
benbrown_ironfoot: lorry-controller-remove-old-jobs moved to lorrycontroller/removeoldjobs.py and simplified somewhat11:58
paulsher1oodbenbrown_: from the original morph functionality... user can have a local directory containing (say) a checked-out component, do work there, cause ybd to build that11:59
benbrown_paulsher1ood: with uncommitted contents though?12:00
paulsher1oodyup12:01
benbrown_how does that affect the cache_key then?12:01
benbrown_or does it not?12:01
paulsher1oodquite. ybd cache-key includes hash of the directory tree - which works for uncommitted contents12:02
paulsher1oodgtristan: how does buildstream handle this case?12:03
benbrown_I see12:03
benbrown_ironfoot: also a thing that vacuums the db every so often, such help to prevent it from getting so inflated12:04
benbrown_it's been a while since I last touched it12:04
ironfootI'll make you the reviewer of my new patches :)12:07
benbrown_s/such/should/12:08
benbrown_heh, ok :)12:08
gitlab-br-botdefinitions: merge request (staging/sam/gcc7.1-bootstrap-fix->master: Fix build issues in master after GCC 7.1 upgrade) #39 changed state ("opened"): https://gitlab.com/baserock/definitions/merge_requests/3912:13
gitlab-br-botlorries: merge request (sam/ccache-tarball->master: Add ccache tarball lorry) #19 changed state ("opened"): https://gitlab.com/baserock/lorries/merge_requests/1912:18
SotKironfoot: neat!12:18
ssam2is anyone able to review that one ?12:22
gitlab-br-botlorries: merge request (sam/ccache-tarball->master: Add ccache tarball lorry) #19 changed state ("opened"): https://gitlab.com/baserock/lorries/merge_requests/1912:22
ssam2our ccache builds actually ran from a baserock/morph branch of ccache.git that had a tarball committed12:22
gitlab-br-botlorries: merge request (sam/ccache-tarball->master: Add ccache tarball lorry) #19 changed state ("merged"): https://gitlab.com/baserock/lorries/merge_requests/1912:22
ssam2thanks!12:22
ironfootxD12:26
ssam2ok, now I can rebuild the world from scratch for the 50th time! \o/12:27
gitlab-br-botdefinitions: merge request (staging/sam/gcc7.1-bootstrap-fix->master: Fix build issues in master after GCC 7.1 upgrade) #39 changed state ("opened"): https://gitlab.com/baserock/definitions/merge_requests/3912:27
ironfootssam2: enjoy the afternoon :)12:27
benbrown_paulsher1ood: Having tested the file:/// logic locally, it doesn't seem to work with uncommitted contents anway13:45
benbrown_anyway*13:45
benbrown_unless there's some magic that I need to put in place of ref?13:45
benbrown_as you checkout by ref, all the code does it say (use the mirror at file:///, rather than cloning it to $git_dir13:47
benbrown_s/it/is/13:47
* benbrown_ is failing to type today13:47
gtristanpaulsher1ood, exactly that way, a recursive hash of the directory; that only applies the an open workspace13:52
gtristanwhich is on a branch and landing soon13:52
gtristanactually there is an MR up for workspaces as of less than an hour ago :)13:52
jjardonssam2: ironfoot Hey! I've created a team in digitalocean so we do not have to use my personal account for the runners; do you want to be part of it? Can you send me an email I can use?14:41
ironfootjjardon: I've /msg'd you my email14:44
ironfootIt's good that more than one person can access to those things14:44
* ironfoot hopes there isn't any maintenance involved14:45
jjardonsure, invitation sent!14:48
jjardonironfoot: we only have to maintain 1 runner, so the maintenance is minimal14:48
ironfootI'm in :)14:51
*** gtristan has quit IRC15:08
*** jude_ has quit IRC15:14
*** gtristan has joined #baserock15:19
*** jude_ has joined #baserock15:31
*** Guest95636 has quit IRC15:44
*** jude__ has joined #baserock15:45
*** jude_ has quit IRC15:45
*** jonathanmaw has quit IRC16:08
*** toscalix has quit IRC16:12
*** jude__ has quit IRC16:28
*** jude_ has joined #baserock16:28
*** lc_ has quit IRC17:00
*** jude_ has quit IRC17:33
*** ssam2 has quit IRC18:58
*** rdale has quit IRC19:08
*** jude_ has joined #baserock19:22
*** noisecell has quit IRC20:37
*** gtristan has quit IRC20:50
*** jude_ has quit IRC21:09

Generated by irclog2html.py 2.15.3 by Marius Gedminas - find it at mg.pov.lt!